Collaborating on Multiparty Information Systems Development Projects

21 downloads 59432 Views 511KB Size Report
Information Systems Group/IOMS, Stern School of Business, New York University, ... qualitative field study of a Web-based application development project was ...
informs

Information Systems Research Vol. 16, No. 2, June 2005, pp. 109–130 issn 1047-7047  eissn 1526-5536  05  1602  0109

®

doi 10.1287/isre.1050.0055 © 2005 INFORMS

Collaborating on Multiparty Information Systems Development Projects: A Collective Reflection-in-Action View Natalia Levina

Information Systems Group/IOMS, Stern School of Business, New York University, 44 West Fourth Street, Suite 8-78 (KMC), New York, New York 10012, [email protected]

G

rowth of Web-based applications has drawn a great number of diverse stakeholders and specialists into the information systems development (ISD) practice. Marketing, strategy, and graphic design professionals have joined technical developers, business managers, and users in the development of Web-based applications. Often, these specialists work for different organizations with distinct histories and cultures. A longitudinal, qualitative field study of a Web-based application development project was undertaken to develop an in-depth understanding of the collaborative practices that unfold among diverse professionals on ISD projects. The paper proposes that multiparty collaborative practice can be understood as constituting a “collective reflection-inaction” cycle through which an information systems (IS) design emerges as a result of agents producing, sharing, and reflecting on explicit objects. Depending on their control over the various economic and cultural (intellectual) resources brought to the project and developed on the project, agents influence the design in distinctive ways. They use this control to either “add to,” “ignore,” or “challenge” the work produced by others. Which of these modes of collective reflection-in-action are enacted on the project influences whose expertise will be reflected in the final design. Implications for the study of boundary objects, multiparty collaboration, and organizational learning in contemporary ISD are drawn. Key words: system design and implementation; outsourcing; management of IS projects; critical perspectives on IT; interpretive research; ethnographic research History: Cynthia Beath, Senior Editor; Michael Myers, Associate Editor. This paper was received on May 22, 2003, and was with the author 16 months for 3 revisions.

Introduction

ness stakeholders, strategists, creative professionals, and brand specialists need to work together, often for the first time, to design and implement new Web-based offerings. Organizations put these diverse professionals on ISD project teams as a means of addressing market necessities or in hopes of igniting creative sparks to discover new market opportunities (Leonard and Swap 1999). Although the potential for producing innovative outcomes is high, the potential for conflict and stagnation is even higher (Guinan et al. 1998, Polzer et al. 2002, Reagans and Zuckerman 2001). This article addresses two questions: (1) how do people from diverse professions and organizational settings collaborate on ISD projects, and (2) how does their diversity influence the systems they eventually design? Insight with respect to these questions will be derived from an ethnographic field study of a Web-

The emergence of the World Wide Web technologies has added several new facets to ISD projects. First, there is a growing involvement of marketing and advertising specialists in ISD projects to help entice consumers to adopt new applications. Second, many new-economy businesses use Web-based applications as the primary means by which to interact with customers and suppliers, hence increasing the strategic impact of these applications. Finally, many of these systems are developed under extreme time constraints and often with the help of external consultants as firms try to gain competitive benefits from being the first movers in the market. These contemporary dimensions of ISD exacerbate the challenges associated with combining diverse forms of expertise on particular projects. In additional to different kinds of technical specialists and busi109

110

Levina: Collaborating on Multiparty Information Systems Development Projects

based application development project. I have adopted a practice-theoretical stance (Bourdieu 1998) and focused on what people do and how their actions shape, and are shaped by, diverse sources of power (relational resources that agents can use to influence their own and others’ actions). Using Schön’s (1983) concept of reflection-in-action, I propose that collaboration on multiparty ISD can be understood as a collective reflection-in-action cycle that shapes and is shaped by diverse organizational and professional stakeholders. This novel theoretical approach, when combined with in-depth empirical research, will extend our current understanding of contemporary ISD settings.

Current Views on Collaboration and Expertise Coordination in ISD

Recent organizational literature has emphasized that effective collaboration among diverse stakeholders enables organizations to draw on diverse forms of expertise to create new competencies and produce synergistic solutions to complex problems (Carlile 2004, Hardy et al. 2005). Although there is little agreement when it comes to defining “collaboration” (Hardy et al. 2005, p. 58)—other than working jointly with others (Merriam-Webster 1998)—IS and organizational scholars have recently clarified what constitutes effective collaboration (Hardy et al. 2005, Hoegl and Gemuenden 2001, Sarker and Sahay 2003). I draw on the Hardy et al. (2005) work to define effective collaboration as a collaboration that leverages the differences among participants to produce innovative, synergistic solutions and balances divergent stakeholders’ concerns. The IS literature provides strong support for the importance of achieving effective collaboration1 in ISD efforts (Kraut and Streeter 1995). Faraj and Sproull (2000), for example, argue that coordination of diverse expertise is a more important predictor of project effectiveness than traditional factors such 1 Although the IS literature often uses the term “coordination” to mean “directing individual efforts toward achieving common and explicitly recognized goals” (Kraut and Streeter 1995), Hoegl and Gemuenden (2001) argue that effective coordination is one critical facet of effective collaboration. Thus, the insights of prior work on coordination are relevant here.

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

as administrative coordination, individual expertise, or development methodologies. One of the key ways by which collaboration is accomplished in ISD is by involving all relevant stakeholders in joint project teams with specific joint deliverables (Iivari et al. 1998). Yet hands-on involvement of diverse stakeholders in design activities does not always lead to effective collaboration and instead may lead to more conflicts (Barki and Hartwick 2001, Newman and Noble 1990, Robey et al. 1993, Sambamurthy and Kirsch 2000). So what do we know about achieving effective collaboration on projects? One of the prerequisites to effective collaboration is the development of a collective identity (Hardy et al. 2005, Kilker 1999) and practices (Levina and Vaast 2005) among participants, through which some of their differences are reflected on and negotiated. On multiparty ISD projects this is accomplished by engaging in a discourse across organizational and professional boundaries (Hardy et al. 2005, Levina and Vaast 2005, Maguire et al. 2004). I draw on the Hardy et al. (2005) notion of collective identity, which, in turn, builds on long-standing traditions in sociology and organizational theory, to conceptualize collective identity as a representational and thus discursive resource that helps collaborators develop a sense of belonging by stressing similarities or shared attributes: Collective identity “names” the group—it gives it an identity that is meaningful to its members and to its stakeholders—and is shared, in the sense that members collectively engage in the discursive practices that produce and reproduce it over time (Hardy et al. 2005, pp. 61–62).

If agents are able to produce and draw on a collective identity as a resource for addressing diverse interests in a balanced way, an effective collaboration is more likely (but not guaranteed) to ensue (Hardy et al. 2005). One of the critical ways by which a collective identity is produced on ISD projects is through participants sharing and negotiating the meaning attached to the artifacts they produce to represent their separate and joint competencies and interests (Sarker and Sahay 2003, Suchman 1995). These artifacts may use case scenarios, functional specification documents (Bødker 1998, Byrd et al. 1992, Kyng 1995), or

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

system prototypes (Carlile 2002). Through the production of these artifacts, ISD participants not only negotiate their relationships and establish a collective identity, but also engage in designing the actual system. The entrance of graphic design, marketing, and strategy professionals into ISD settings adds to the challenges for establishing effective collaboration. First, these professionals bring with them their own distinct, professional interests, which can make the process of developing a collective identity more complex. Second, these professionals tend to be more focused on the interests of users as compared to business stakeholders or technical developers. This focus may come into conflict with the interests of other business stakeholders such as production, human resources, and information technology (IT) folks, who may resist changes needed to better serve consumers and business partners. Third, new entrants into the ISD scene may have a hard time relating representations of their practices (e.g., market analysis documents) to the representations used by traditional ISD participants (e.g., structured requirements documents). In lieu of mutually familiar design representations, prototypes become increasingly important (Bechky 2003, Carlile 2002, Schrage 2000). Yet there are well-known conflicts associated with prototyping—such as premature convergence on a particular design—which may be exacerbated when more stakeholders and developers with diverse interests become involved (Terwiesch and Loch 2004). Additional problems with establishing effective collaboration in modern ISD settings may stem from the need to coordinate efforts across organizational boundaries (Hardy et al. 2005). It is traditional to assume in the IS literature that clients’ interests should be dominant on projects and that clients should exercise various modes of control to ensure this (Choudhury and Sabherwal 2003, Kirsch et al. 2002, Sabherwal 1999). It is not clear, however, how these recommendations apply to situations when consultants are hired not only to implement the system, but also to recommend a new business direction about which clients have not previously thought and which may meet with organizational resistance. When clients exercise excessive control, consultants may be forced to produce quick, tangible results or say what

111

clients want to hear, thus pushing the collaborative process away from the risk taking needed to develop innovative, synergistic solutions (Clegg et al. 2004). At the same time, however, clients’ attempts to exact control may have limited impact (Gable and Chin 2001), as consultants are sometimes able to use their specialized expertise to control the decision making process on the project (Elkjaer et al. 1991). We need to develop a deeper understanding of the actual collaborative practices on multiparty ISD efforts. Such understanding will shed light on how divergent professional and organizational stakeholders’ concerns can be balanced to produce innovative, synergistic solutions. In what follows, I propose a collective reflection-in-action perspective as a useful lens through which to investigate how diverse participants collaborate on joint projects. This perspective was developed in an inductive, grounded fashion from case study data and only then, was it integrated with existing theories (Glaser and Strauss 1967).

Collective Reflection-in-Action View Reflection-in-Action “Reflection-in-action” was proposed by Schön to understand how individual professionals address uncertain and nonroutine, yet repetitive, problems in practice (1983, p. 60). ISD researchers have used this perspective in investigating individual designer’s practices (Andersen 1990)2 and, to some extent, the user-developer interaction (Bødker 2000). Reflection-in-action is a reflective “conversation with the material of a situation” (Schön 1983, p. 79). The notion emphasizes that “the unique and uncertain situation comes to be understood through the attempt to change it, and changed through the attempt to understand it” (Schön 1983, p. 131). The process of reflection-in-action is a cycle of appreciation, action, and reappreciation (Schön 1983, p. 132). Appreciation consists of thinking and framing the phenomenon. Action consists of experimentation 2 Andersen’s (1990) professional work practice approach has been criticized for understanding ISD as a process of modeling the reality surrounding designers rather than actually shaping it (Iivari et al. 1998). In response to this criticism, I extend Schön’s work to illustrate how actions taken by diverse stakeholders shape their interaction and the reality around them.

112

Levina: Collaborating on Multiparty Information Systems Development Projects

or actions taken to see and evaluate their outcomes (Schön 1983, p. 145).3 A crucial part of reappreciation is the use of the professional’s appreciative systems (perceptions, values, and skills) to judge the outcomes of the action and to guide further experiments (Schön 1983, p. 135). Reflection-in-action is essentially a structurational, practice-based concept (Bourdieu 1977, Giddens 1984). It highlights the duality of structure and action: An agent “shapes” the situation and the situation “talks back” by producing unintended consequences giving the situation a new meaning (Schön 1983, p. 79, 131). In Schön’s view, the “back-talk” can be literal, as in the case of a psychotherapist listening to a patient, or figurative, as when an architect evaluates a drawing that he or she created. An agent’s key concern revolves around the question of which experiments to preserve for the future from among the experiments that produced the present. Although Schön’s work is focused primarily on individual practice, he acknowledges the challenges involved in the reflection-in-action of multiple professionals (1983). In such cases, reflection-in-action involves “on-the-spot surfacing, criticizing, restructuring, and testing of intuitive understandings of experienced phenomenon” (Schön 1983, p. 241). In this paper I expand on Schön’s ideas to describe professional practice in collaborative settings that combine diverse expertise. I introduce the term collective reflection-in-action (CRIA) to describe a “conversation” among agents that brings about dilemmas stemming from differences in agents’ appreciative systems. In CRIA, appreciation and action may be undertaken by one agent while reappreciation and subsequent action are undertaken by another. There is another important difference between individual and collective reflection-in-action. While individual reflection-in-action often involves silent reflection, CRIA necessitates that participants express and share their views explicitly—through the production of audible or visual artifacts—so that collaborators can reflect on their partners’ reflection-in-action. In this way, collaborators engage in discourse. If collaborators do not express their ideas at all (producing 3 Note that the word “experimentation” in this context does not imply novelty of the action.

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

“organizational silence,” according to Morrison and Milliken 2000), collaboration will become impossible. The question of how participants express their ideas through explicit objects is closely related to the concept of boundary objects (Star 1989). Boundary Objects The term boundary objects was introduced in studies of collaborative problem-solving activities undertaken in scientific communities (Star and Griesemer 1989, Star 1989). It refers to “objects that are plastic enough to adapt to local needs and constraints of the several parties employing them, yet robust enough to maintain a common identity across sites” (Star 1989).4 Boundary objects include physical product prototypes (Bechky 2003, Carlile 2002), design drawings (Bødker 1998), shared IT applications (Briers and Chua 2001, Pawlowski and Robey 2005), engineering sketches (Bechky 1999, Henderson 1991), standardized reporting forms (Bowker et al. 1996, Briers and Chua 2001, Carlile 1997, Star and Griesemer 1989), and even shared abstract constructs such as product yield (Kim and King 2000). The growing literature on boundary objects has been primarily concerned with classifying their types (Star 1989) and properties (Bechky 2003, Carlile 2002). For example, it has been argued that to be effectively used for boundary spanning, an object should be tangible (Carlile 1997), concrete (addressing specific problems in a specific community; Bechky 2003, Henderson 1991), accessible (Carlile 1997), and current (Carlile 1997). System prototypes are among the most cited examples of “effective” boundary objects because they are seen as useful means for participants (e.g., marketing and engineering professionals) to represent and negotiate differences in their practices and interests (Bechky 2003, Carlile 2002). Yet many explicit objects do not become boundary objects in practice, as agents do not see their local usefulness or fail to establish a common identity for them across sites (Levina and Vaast 2005). Levina and Vaast (2005) argue that 4 While the term “object” is not defined in the literature on boundary objects, I draw on practice theory to define objects as symbolic representations of practices. Moreover, common identity of an artifact refers to an artifact serving as a common symbolic resource for diverse agents. These ideas are further developed in Levina and Vaast (2005).

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

a common identity for an artifact can only emerge when agents engage in a common set of practices and develop a collective identity (represented through boundary objects). Combining the insights on boundary objects with the CRIA lens can help us further explore the nature of multiparty collaborative practices. Not all objects produced and shared through CRIA become boundary objects in practice (e.g., an object produced by one agent may be useless to another). We need to understand what constitutes effective collaboration by unpacking how agents produce, share, and use explicit objects through CRIA. Collective Reflection-in-Action in ISD Before this investigation is undertaken, note that the CRIA lens proposed thus far does not assume that agents are involved in an ISD practice. Here I draw on practice theory (Orlikowski 2000) to propose that ISD project participants are primarily concerned with designing future IT use practices. Drawing on Orlikowski (2002, p. 256), I define IT use practices as a set of recurrent interactions among agents’ using and modifying an IT artifact. Such interactions are situated in a social and historical context and are bounded by physical surroundings and technological artifacts. Explicit objects produced by ISD participants (the IT artifact, system documentation, orally expressed ideas, use scenarios, etc.) are primarily used to help represent the envisioned IT use practices (Bødker 1998). I refer to the constantly changing, envisioned IT use practices represented through these objects as an IS design.5 At present, IS researchers have a somewhat limited understanding of the collaborative practices through which IS designs are produced and negotiated on multiparty projects (Robey 1994). To deepen our understanding in this area, I engaged in an empirical study of such practices. An ethnographic approach was used to understand this socially complex and embedded phenomenon. Understanding emergent patterns of CRIA required at least a partial reliance on direct observations of practice so as to record who 5

The design that emerges during ISD process is often quite different from the enacted IT use practices once the system is implemented (e.g., Orlikowski 2000).

113

interacted with whom, when, where, and using which objects. Retrospective accounts and archival data were insufficient for these purposes because of recall bias and a lack of documentation. I followed an ISD team comprised of diverse professionals from an Internetconsulting company (Eserve) and a large division of a publishing company (Pubco) engaged in designing a website for Pubco’s customers.6

Research Methods Sites Eserve. In the late 1990s, Eserve was a young, rapidly growing, professional services firm engaged in end-to-end production of Web-based applications. The company was both financially successful and popular with investors, eventually receiving a 90% client satisfaction rating on the basis of a thirdparty survey. In the mid-1990s, Eserve started to provide technical Web-development services. Soon after, strategic consulting expertise was added by hiring management consultants and business analysts. In 1997, responding to competition from design and advertising agencies, Eserve added graphic designers and brand specialists, who found the digital medium, high salaries, and stock options appealing. To integrate “technology,”7 “strategy,” and “creative” disciplines, Eserve leaders promoted an egalitarian culture and “open space” environment. According to the CEO, Eserve’s ability to integrate diverse disciplines was the source of the company’s competitive advantage: To build digital businesses, to get ideas, and to get them launched in the marketplace you need to put three kinds of people [technologists, strategists, and graphic designers] together in a 12 × 12 [feet] workspace. Culture is the key—collaboration, sharing, mutual respect. Real innovation is at the intersection of disciplines (excerpt from a public speech). 6 Names of organizations and their members, titles, products, and technology applications have all been disguised. They have no relationship to similarly named organizations that might exist in the real world. 7 The quotation marks are used in this and the following sections to denote terms used by Eserve and Pubco study participants.

Levina: Collaborating on Multiparty Information Systems Development Projects

114

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

A critical characteristic of Eserve projects (as they evolved by 2000) was their three-phase service delivery model: a planning phase led by strategists with a few graphic designers and technologists on hand; a prototyping phase, which involved more graphic designers and fewer strategists; and a development phase led and staffed primarily by technologists. The so-called “waves” service delivery model is captured in Figure 1. The model specified that strategists, who typically played the roles of engagement managers (referred to as “client partners”), project managers, and business analysts, were to interface heavily with clients and make key decisions, while technologists and graphic designers were to be primarily involved in tasks associated with building the website. The objective of the business development stage carried out at Eserve’s expense was to negotiate a fixed-fee deal for one or many phases of the project. The planning phase consisted of one-on-one interviews and facilitated workshops with the client as well as market and strategy analyses intended to identify a strategic direction for a client’s Web offering. The prototyping phase focused on defining system requirements and a mockup prototype for the website. In the development phase, the actual functionality was implemented. Each phase was contracted separately upon the completion of the prior phase. Pubco. Pubco was the educational book division of a multidivisional publishing company. Its strong hierarchical and departmental distinctions

Degree of project involvement

Figure 1

The “Waves” Service Delivery Model

Strategists

Graphic designers

Time

Technologists

Planning phase

Prototyping phase

Development phase

were characteristic of the industry (Epstein 2001). Authors and illustrators dealt primarily with the editorial staff, while consumers and booksellers dealt with those in sales and marketing. Within its industry, Pubco saw itself as a firm focused more on high-quality books than on sales volume. Pubco’s IT department played only a supporting role (mostly for the finance and sales groups), while several decentralized Web-development teams supported the editorial staff. Pubco’s existing website had been built without any overarching market or brand strategy. By fall 1999, Pubco’s top executive—under pressure from sales and marketing—concluded that an integrated website was needed to reach more of Pubco’s customers and enhance Pubco’s brand. In September of 1999, Pubco began looking for a partner for the project, eventually picking Eserve. The planning phase started in January 2000. Data Collection and Analysis Starting in January 2000, I conducted an eight-month field study of the Eserve-Pubco project using primarily qualitative data-collection techniques (Agar 1980, Barley 1990, Glaser and Strauss 1967, Schwartzman 1993, Van Maanen 1979). I negotiated my entry into the project on Eserve’s side after observing the company’s new employee training program. As a nonparticipant observer, I spent four to five days a week, six hours a day, in different settings at Eserve and Pubco, recording my observations in 20 plus typewritten notes each day. I observed the Eserve-Pubco project every week starting in the second week of the planning phase and every other week during the prototyping phase. However, I did not follow the development phase because my analysis had reached the stage of theoretical saturation, where no new analytical concepts were emerging from the data (Glaser and Strauss 1967). I did, nonetheless, make several follow-up visits when the application was launched. Day-to-day observations took the form of attending key project meetings and shadowing specific individuals, primarily at Eserve. While I was slightly limited in my analysis by the lack of daily observations at Pubco, I felt that spending time with Eserve participants allowed me to observe more events directly tied to the design’s evolution.

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Detailed observations were supplemented with 19 in-depth interviews at Eserve and 21 at Pubco.8 The interviews were conducted at the end of the planning phase and again at the end of the prototyping phase of the project and involved all project participants (see Table 1 for participants’ roles). Those participants involved in both phases were interviewed twice, and each interview was tape recorded and transcribed verbatim. To incorporate the wider organizational context and the backgrounds of participants into the analysis, I obtained supplemental data from Eserve’s human resources database, Web pages, industry reports, popular press accounts, and financial statements. In addition, I analyzed the project’s e-mail, printed documentation, and Intranet-based documentation archives. I also took digital pictures of many flip charts, whiteboard drawings, and other objects. To understand the nature and evolution of collaborative practices and the way they shaped the IS design, I have conducted two analyses. The first analysis was motivated by the boundary objects concept.9 This analysis focused on understanding how interaction among designers shaped the IS design. I examined my notes, interview transcripts, and documents to uncover the chronological evolution of the design from the legacy Web-based system, through early idea generation and subsequent design choices, and, eventually, to the prototype. I treated each new idea about the emergent system as a version of the system design and represented these ideas in a graph-like structure (following Miles and Huberman 1984). While conducting this analysis, I became increasingly sensitive to the utility of Schön’s work and subsequently developed CRIA. During the second analysis, I looked at how the differences in participants’ backgrounds and practices exhibited themselves in the enactment of CRIA and shaped the project’s discourse. I looked at 8 Twenty additional interviews were conducted at various Eserve’s corporate groups not directly involved with Pubco. 9 The analysis is new to the literature on boundary objects. Prior studies of boundary objects (reviewed above) were concerned with the properties, types, and roles of these objects, not with how the use of an object shaped the resulting product. Thus, I had to develop a new analytical technique to address my research question.

115

institutionalized sources of difference such as interorganizational, professional, departmental, age, gender, ethnicity, education, organizational status, as well as organizational and professional tenure distinctions (Polzer et al. 2002). While all of these played some role with respect to the project, the interorganizational and professional differences were most influential in shaping the IS design. I then evaluated the data that was not explained by the differences I had previously considered. It became apparent that a new distinction, stemming from the intensity of participants’ involvement in influencing the IS design (projectbased diversity), enabled the participants to negotiate a collective identity while also serving as a source of new tensions.

Research Results

As suggested by prior research on ISD, the design of the Eserve-Pubco website involved explicating design ideas either through references to existing artifacts (Pubco’s existing website) or development of new artifacts. Yet, after examining hundreds of objects used on the project, I concluded that whether an object was used at all and whether its use enabled or impeded effective collaboration could not be understood through the stable properties of the object itself, but only through the use of the object in practice. Thus, I developed and used CRIA to explain these diverse uses of objects in practice. The use of an object in practice could be classified as one of three modes of CRIA enacted by participants: ignoring, adding, and challenging (see Figure 2 for an illustration). In the “ignoring” mode, an explicit object produced as a result of Agent A’s (virtual or physical) experimenting with the design is not reflected upon (no sense is made out of it) in Agent B’s subsequent experimentation. Ignoring may happen as a result of not receiving, not mentally registering, or not understanding the object. In the “adding” mode, an object produced as a result of Agent A’s experiment with the design is reflected upon by Agent B and preserved in his or her subsequent experimentation. Finally, in the “challenging” mode, Agent B reflects on the object produced by Agent A but eventually generates an object that does not preserve all of Agent A’s design. Participants sometimes referred to this mode of work as undoing.

Levina: Collaborating on Multiparty Information Systems Development Projects

116

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Table 1

Team’s Composition and Deliverables

Phases

Business development

Planning

Prototyping

Duration

October ’99–January ’00 3.5 months

January ’00–March ’00 7 weeks

March ’00–June ’00 13 weeks

Eserve team composition

Client partner (Frank) 1 Sr. Strategist (Van) 1 Jr. Strategist (Adam) joined after the sales presentation 1 Jr. Technologist (Boris)

Client partner (Frank) Project Manager (Bob) 3 Jr. Strategists (Adam, Cherry, Nicole) 1 Jr. Technologist (Boris) 1 Sr. Technologist (Kirk) 2 1 Brand Specialist (Henry) 2

Client partner (Frank) Project Manager (Wendy) 2 12 Jr. Strategists (Adam, Cherry, 1 time Nicole) 2 1 Jr. Technologist (Boris) 1 Sr. Technologist (Kirk) 2 Jr. Graphic Designers (Dolly, Jason) 1 Jr. Information Architect (Dick) 1 Brand Specialist (Henry) 2

Pubco core team composition

Top Executive (April) Finance VP (John) 4 Sales and Marketing Mgrs. (Keri, Todd, Lily, Jen) 1 Editorial Mgr. (Joan)

Finance VP (April) Pubco Project Manager (Maya) 2 Sales and Marketing Mgrs. (Keri, Lily) 1 Editorial Mgr. (Joan) 1 Print Production Mgr. (Sarah) 1 IT Mgr. (Sally) 2 Web Group Mgrs. (Alice, Natalie)

Finance VP (April) Pubco Project Manager (Maya) 2 Sales and Marketing Mgrs. (Keri, Lily) 1 Editorial Mgr. (Joan) 1 Print Production Mgr. (Sarah) 2 IT Mgrs. (Sally, Brian) 2 Web Group Mgrs. (Alice, Natalie)

Deliverables

Sales presentation Agreement on scope and price (statement of work) Contract

Customer segmentation analysis Customer interview report Market trend analysis Initiative selection Functionality prioritization Assessment of existing technological initiatives Technology platform selection

Use cases New site map Style guide Click-through mockup Prototype of one functionality Technical architecture Usability test report

These three modes of collaboration are mutually exclusive and exhaustive: If an object is reflected on, then an agent’s appreciative system changes, and this change affects subsequent actions, which can either preserve or transform the design. Moreover, it is Figure 2

Three Modes of CRIA

9.00 AM

9.30 AM

Ignoring

Agent A Adding Challenging

10

Challenge mode Add mode Ignore mode

Agent B

possible that at certain times in the design process, collective reflection-in-action coincides with individual reflection-in-action. In other words, the person involved in producing subsequent objects is the same person who produced the original object. Given the definition of effective collaboration offered here, it is critical that, at some point, the challenging mode of CRIA be enacted, so that agents either challenge their own prior thinking about the design (on reflection on the other agent’s objects) or challenge objects produced by another agent. Should this not happen, innovative, collaborative solutions will not be possible, as each party would ultimately continue to do what it had done before without “transforming” its expertise in practice (Carlile 2004).10 (See Figure 3.) One may argue that a synergistic solution can be created through the adding mode of practice alone. However, in the ISD practice, as in any new product development practice, we can assume that there are always areas of ambiguity and complexity that a mere combination of diverse skills, without their “transformation,” cannot address (Carlile 2004).

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Figure 3

Agents Shaping the IS Design Through CRIA Agent B reflects and adds

Agent A reflects and experiments Agent C reflects and challenges

Next I provide an overview of the evolution of the Eserve-Pubco project using examples of objects that were produced and shared by participants (a more detailed account of each phase is provided in Levina 2001). An overview of each stage of the project is followed by the CRIA analysis. The following subsection explores why different modes of CRIA were enacted in particular situations by focusing on how the organizational and professional diversity among participants shaped the design.11 Eserve-Pubco Project Evolution and CRIA Analysis Business Development Stage. In September 1999, Pubco started looking for Internet consulting firms to quickly revamp its website, which included more than 100 pages with inconsistent styles, poor use statistics, and often no references to Pubco’s brand. We really innately do not understand the Web, and I think we are trying to figure out how we can participate in this new environment quickly. Of course, we are also trying to figure out how we would make money on it, but that part is not yet evident (one of Pubco’s executives).

Several factors motivated Pubco to favor Eserve’s sales pitch, including a sales meeting involving a senior Eserve strategy consultant who was experienced in Eserve methodologies and in Pubco’s industry. Following a review of Pubco’s and its competitors’ Web strategy, the consultant proposed several creative ideas that Pubco’s managers liked. These managers were also impressed with Eserve’s credentials in brand strategy and graphic design. Eserve proposed a thorough redesign of the existing website 11

Due to space limitations, technologists’ role in the project will be skipped (see Levina 2001 for a more detailed account).

117

and the development of extensive new functionality and accompanying marketing strategies. The specific recommendations were put forward in a sales presentation given to Pubco’s top managers by the senior strategy consultant. As time went by, however, it became clear to Eserve that Pubco was unwilling to spend the amount of money Eserve customarily charged for a team of top-notch professionals. The senior strategy consultant was then assigned to a different project, and, in the statement of work included in the contract and signed by representatives of both organizations, the project was scaled down to a comprehensive website redesign with the relatively modest strategic goal of addressing the needs of an underserved group of consumers. Rather than six to nine weeks, the planning phase was shortened to six. Despite these changes, Pubco’s top managers reported that although they could not spend as much as some of Eserve’s other clients, they were expecting to get Eserve’s best consultants and to see innovative results, especially if Eserve were to leverage Pubco’s existing strategic and marketing analysis and Web-development efforts. This approach that was atypical for Eserve projects. Eserve’s branch manager put together a team consisting mostly of junior consultants. Although Eserve promised all its clients that 60% of the consultants on each team would have more than six months of Eserve experience, the company’s high growth rate made this challenging. As such, the Pubco project was staffed with only 23% Eserve-experienced consultants. At Pubco, a top executive hand picked a “core team” of seven midlevel Pubco managers, plus a vice president as the project sponsor, and a market research consultant as their project manager (Table 1 depicts each team’s composition as the project progressed). According to their personnel records, the two teams were quite different. The Eserve team members were younger (the average age was 30 as opposed to 44 at Pubco), predominantly male (6 men and 2 women at Eserve compared to 2 men and 7 women at Pubco), and likely better compensated than Pubco participants.12 12

According to the U.S. Bureau of Labor Statistics (www.bls.gov/ oes), the mean annual salary in the computer systems design industry was 65% higher than in the publishing industry, totaled across occupations.

118

Levina: Collaborating on Multiparty Information Systems Development Projects

Despite initial enthusiasm, Pubco approached the new project, which was about to cost 6% of Pubco’s operating income, cautiously: I tried to make it clear to Eserve that if they do not get an understanding quickly of who we are and how we do business, they would fail just like other consultants with whom we had worked had failed (Pubco executive).

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Eserve’s ability to deliver. Much of the knowledge about Pubco’s business gained by Eserve during the business development stage had been lost when the senior strategy consultant left and several newcomers joined the project in the planning phase (see Table 1): Excerpt from field notes Week 2 of the project: I am following Nicole—an Eserve strategy consultant who joined in the planning phase:

CRIA Interpretation of the Business Development Stage. During business development, Eserve reflected on Pubco’s prior work and on Eserve’s own expertise in Web strategy and design to generate what all participants believed to be innovative design ideas captured in the sales presentation. In CRIA terms, Eserve consultants engaged in the challenging mode of CRIA in producing the sales presentation: They reflected on Pubco’s existing technical and strategic initiatives and suggested how to both build on what had been done in the past and “undo” what was not working in the present. The sales presentation was, by all accounts, an effective boundary object during the business development stage, as its use suited Pubco’s and Eserve’s needs and helped build a collective identity for the project: It was concrete (it referred to the problem at stake), tangible, and accessible to other project participants (printouts were made for all participants). The use of visual language helped make key points. While Pubco’s participants signed a contract that pointed toward a greater reuse of (adding onto) Pubco’s existing objects in subsequent design activities, it was Pubco’s hope that Eserve would continue to creatively challenge Pubco’s prior work so as to ensure the development of the ideas captured in the sales presentation.

As this example demonstrates, instead of reflecting on what other Eserve consultants had learned during business development, the newcomers were working to meet deadlines on day-to-day tasks on the project, often reinventing the wheel. Pubco’s reaction to this approach was not surprising:

Planning Phase. A week after the planning phase began (before I started observing the project), I met Eserve’s project manager (Bob). He was excited about my joining the project because he believed the relationship with the client was strained. Bob welcomed the opportunity to reflect on the relationship with me especially given that he was struggling with the client’s project manager (Maya), who was insisting on a lot of seemingly unnecessary documentation that would demonstrate Eserve’s comprehension of Pubco’s existing strategy and marketing documents. As I started observing day-to-day events directly, it became clear to me that Pubco was questioning

I was continuously disappointed in organizational things, the ability to connect the work, the documents that were given to them [Eservers], for example, with the work that comes out     We made a point and spent some time on developing a list of interviewees [within Pubco]—people they should talk to at the beginning of the project. It seems like that got thrown away     No interview write-ups [for interviews with business stakeholders that took place], no analysis of past projects, and documentation of having read them. And four weeks into the project, we discovered that they had lost a copy of the strategic plan we gave them. So, we gave them another one. Well, that was just sort of a little shocking to us (Pubco team member).

Bob (after a meeting at Pubco): They want us to do a lot of prep before we go into a meeting with them. That would be hard to balance because after we do prep and talk to somebody, they also want us to do more. Nicole: Do we have notes from the sales meeting? Adam (strategy consultant involved in business development): Isn’t it in this folder [points to the folder on his desk]? Later in the day, Nicole is explaining to me that she is going to work more on the interview guide [for interviewing Pubco’s unit managers]. I asked if she plans to read the documents from the pile [pointing to the documents from business development phase]. She said she would take them home with her. She said that she has to be done with the interview guide today to send it to Clare [a manager at Pubco], “whoever she is.” Nicole said that reading the documents might influence the interview guide, but that that’s okay [implying that she would send the guide without reading the documents].

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

As the third week of the planning phase approached—a time when, according to the Eserve service delivery model, consultants should be up to speed on their client’s business—Eserve conducted a brainstorming workshop with the intention of generating ideas as to what kind of functionality the website should support. Using Eserve’s brainstorming methodology, Pubco participants generated more than 100 ideas for new site functionality, many of which radically departed from Pubco’s current Web strategy and organizational practices. Eserve consultants played more of a facilitative role during the workshop because they did not yet feel comfortable contributing their own ideas while still learning about Pubco’s business. After the workshop, the initiatives were consolidated into themes by Eserve strategists and then discussed in a subsequent workshop with Pubco, which further clarified the meaning of the initiatives to participants. The next step was for Eservers to rate this consolidated list of initiatives on the basis of 15 criteria drawn from Eserve’s methodology (examples of criteria included technical feasibility and strategic impact). The relative weight of each criterion in the initiative selection was provided by Pubco participants; however, the scoring was to be done by Eserve based on consultants’ research and Web-based expertise. The entire fourth week of the project was allocated to this exercise. During this time, consultants picked up steam. Not only did they conclude most of their research—including the analysis of Pubco’s existing technology initiatives, industry and competitive landscape, and interviews with consumers (potential and current users)—but they also sought guidance from Eserve’s senior strategy experts. Meanwhile, two junior strategy consultants became worried that the Eserve team was not holding its own brainstorming session and was therefore missing an opportunity to generate “truly innovative” Eserve ideas (such sessions were not planned because Eserve’s methods assumed that consultants, who would have had more time to prepare during the normal 9- to 12-week planning phase, would generate initiatives in the brainstorming workshop along with clients). Eservers then held an internal brainstorming workshop, thereby generating several initiatives that were in some ways similar and in other ways dif-

119

ferent from what their clients had already proposed. Eservers then added these initiatives to those generated by the clients. It was during this fourth week that clients decided to deviate from Eserve’s planning phase plan. Struggling to understand Eserve’s methodology and somewhat suspicious of Eserve’s grasp of Pubco’s business, Pubco core team participants from sales and marketing became worried that the initiatives they believed to be essential had been overlooked during the brainstorming session (while these initiatives were discussed during brainstorming, Eservers assured Pubco that these initiatives were so basic as to not warrant recording). Maya conducted two Pubco-only workshops where core team members identified and prioritized their “critical” initiatives and presented them in a document entitled “Key Elements of Our Web Presence that We Know We Need to Develop for Our Core Business.” This “must-have” list was relayed to Bob and sent by Maya to all consultants to be included with the previously brainstormed initiatives: Excerpt from field notes, 2-15-00 I am attending a meeting at Pubco: Maya (addressing Bob): We don’t have a lot of functionality that we absolutely need. We are not trying to stop the process. We are trying to make sure that you [Eserve] know these are the things we are going to be looking for quickly     If we put it through our prioritization matrix, we will end up with them.

Consultants received the must-have list as they were preparing to embark on their prioritization exercise. They included the list in the prioritization process along with the initiatives they and the Pubco people had brainstormed previously. As Eservers started rating the initiatives, it became clear to them that Pubco’s must-have initiatives were not getting high priority. For example, many of these initiatives were of no interest to the end users. At the same time, Eservers felt that Pubco was “too traditional” to understand the value of the more innovative initiatives brainstormed on the project and that Pubco was putting pressure on them to recommend the must-have list. In addition, because of their lack of experience at Eserve, newer consultants themselves were struggling to understand how the prioritization process was supposed to work. To save the relationship and obtain funding for the prototyping phase, Eserve consultants decided

Levina: Collaborating on Multiparty Information Systems Development Projects

120

Figure 4

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Evolution of the IS Design Through Eserve-Pubco Interaction

Business Development

Objects Previously Produced by Eserve

Objects Previously Produced by Pubco Objects produced through CRIA across groups Existing Website Strategic Documents Organizational Processes IT Architecture & Market Analysis

Planning Phase

Web Strategy and Design Methodologies

Sales Presentation

Planning Phase Contract

Pubco Brainstormed Initiatives

User Interviews & Interviews with Pubco Managers

Consolidated Themes of Initiatives

Eserve Brainstormed Initiatives

Must Have List

Competitive Analysis, Feasibility Analysis

Recommended Initiatives

DILO

Prototyping Phase Pubco Use Cases

Challenge Mode

Time

Pubco Wire Frames, New Site Map, Mockup

to abandon the formal prioritization process. In the client meeting, Eservers presented several sets of initiatives originating from the two brainstorming sessions and from the must-have list. They then developed justifications and produced ratings supporting why Eserve was actually recommending the musthave list initiatives. During my interviews, several of Pubco’s participants expressed disappointment with Eserve’s recommendations: In some ways these are ideas [initiatives that were ultimately chosen] that a bunch of people here and I have had for a while, but you have to hire a consultant and pay them a lot of money to make sure that the ideas are comfortable to everybody     And now they [Eserve] are only listening and not pushing back. I have never felt like they really pushed at us. Well, you know, that is hard. Maybe I am having a hard time evaluating them now; maybe they pushed and I just did not like it, you know. I do not know. Um, when you push someone, you need to provide, you know,    an alternative (Pubco’s core team member from the sales and marketing function).

Add Mode Ignore Mode

Although nobody at Pubco thought the recommended initiatives were innovative, Pubco’s top executive did find a multistep graphic depiction of a potential user engaging with the proposed website’s features particularly helpful in understanding and supporting the proposal. Eservers referred to this storyboard-like tool as “a Day In the Life Of” (DILO) a user. While it was increasingly clear that Eserve’s project was not resulting in innovative strategic ideas, Pubco participants believed the DILO was indicative of just how valuable Eserve’s graphic design expertise really was. In fact, Pubco’s top executive took the DILO to an executive retreat to share it with the leaders of other divisions: I think what Eserve brought to it is a way of communicating these very complex issues in a specific or applied way, as opposed to just in a theoretical way. You can talk to me until you are blue in the face about communities or about business to business on the Web or whatever, but they did a fabulous job of taking all the things we have been working on and laying them out in this sort of diagrammatic way (top executive).

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Looking back at the planning phase of the project, she continued: I think everyone always felt Eserve people were smart, but they were not always getting the feedback, or even understanding anything we were talking about, or internalizing it, using it, and then forming their opinions or judgments. So, something happened along week three and a half or four that turned this around.

CRIA Interpretation of the Planning Phase. The design’s evolution—as a result of Eserve and Pubco’s collaboration—is illustrated in Figure 4. It is notable that many objects produced separately or jointly by Eserve and Pubco were ignored in subsequent iterations of the design. For example, the business development phase sales presentation (a tangible, concrete, and accessible object, available to all participants) was ignored in the planning phase. Also, Pubco participants ignored Eserve’s methodology when it came to prioritizing the initiatives. On the other hand, there were cases of creatively challenging the design. For example, by the fourth week Eserve strategists were able to better relate to Pubco’s practices and interests and were willing to challenge Eserve’s established process by conducting an internal brainstorming session to deliver a more innovative solution for Pubco. Finally, the DILO object was produced by adding to the recommended initiatives object by drawing on Eserve’s Web design methodologies. Prototyping Phase. During the prototyping phase, following Eserve’s “waves” model, three new design discipline members (two junior graphic designers and an information architect) and one senior technologist joined the team. It was expected the graphic designers would deliver the graphic treatments for the website mockup—a click-through prototype with minimal functionality. Another major change in this phase was the arrival of Wendy, a new Eserve project manager who was a business analyst (a strategist, according to Eserve’s classification) and who had extensive experience managing Web-based projects. Also, after much discussion among managers on both sides, it was decided that to improve Eserve-Pubco communications the team would be divided into three smaller integrated subteams (requirements, design, and technology). Wendy and Maya would attend meetings of each subteam as liaisons. Table 1 lists participants in this phase.

121

When the new members officially joined the team, they were given copies of deliverables from the planning phase. In a “knowledge dump” meeting, the planning phase strategists walked the newcomers through the planning phase deliverables presentation containing the DILO. In addition, the newcomers were encouraged to browse the existing Pubco site. The graphic designers, who did not know the publishing industry well and had limited interactions with the client, had difficulty understanding the DILO and existing website terminology. Wendy, on the other hand, was continuously interacting with Pubco participants in various meetings and was able to understand Pubco’s language fairly well. Meanwhile, Eserve strategists (Adam and Cherry) and Pubco managers assigned to the requirements subteam (Kerri, Lily, and Joan) were building “use cases.” Use case scenarios are part of the universal modeling language (UML) object-oriented systems analysis methodology and typically rely on graphic representations of users’ and systems’ actions (e.g., Jacobson 1993). However, Eserve strategists were not familiar with UML; thus they decided to create “Pubco use cases” in the form of traditional functional requirements captured as hierarchical bullet lists of text. Eservers produced the initial Pubco use cases based on their market research (e.g., what Amazon has done), planning phase documents, and Pubco’s current website. Then members from both organizations met to debate Pubco’s use cases. Requirement team members from both organizations expressed satisfaction with what they saw as creative interactions in the requirements’ meetings, and project participants believed that innovative functionality emerged as a result of these interactions. Once the use cases were finalized, they were passed on to Eserve’s graphic designers (to develop a website mockup) and the information architect (to produce a new site map). Weeks were spent in meetings explaining the envisioned functionality to graphic designers who were still “not getting” Pubco’s project. Eserve strategists grew increasingly frustrated that the graphic designers were not “taking ownership” of the website and were not “making an effort to learn.” By the middle of the prototyping phase, the designers were only able to produce a color palette and a font

Levina: Collaborating on Multiparty Information Systems Development Projects

122

Figure 5

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Evolution of the IS Design Through Strategist-Designer Interaction

Planning Phase Objects Previously Produced by Strategists

Objects produced through CRIA by Designers and Technologists on the project

Structured Bullet Lists Ranking Matrices Textual Analysis Documents

Prototyping Phase

Objects Previously Produced by Designers Story-boards/Use Scenarios, Drawings, Site Maps Competitors Websites Usability Guidelines Pubco's Existing Website's Graphic Design

DILO

Pubco Use Cases

Existing Website Map

Pubco Existing Website Functionality

Content - free Mockup Pages Color Palette and Font Guide

Wire Frames Complaints On the Number of Navigational Elements New Website Map

Graphical Treatments for the Mockup

Extra Mini Design Project Time

Challenge Mode

Redesigned Graphical Treatments

Add Mode Ignore Mode

guide (based on their review of Pubco’s current website) and one content-free (but color) mockup page. The graphic designers, however, were trying to learn Pubco’s business. In their attempt to do so, they produced a new inventory of Pubco’s existing website in the form of a tree-like site map. This object was easily interpreted by all participants. Despite this, the graphic designers and the information architect were still unable to produce a navigational structure or an index page for the new website. Toward the midpoint of this phase, with deadlines passing one after another, Wendy began reflecting on the way in which strategists were communicating with the graphic designers. Following a number of contentious conversations with the designers, Wendy realized the designers were not accustomed to working with structured documents such as Pubco use cases and were not using the documents that had been sent to them. One designer openly admitted in a meeting that he had no interest in learning

to read structured documents (i.e., Pubco’s use cases) and that he had disposed of the printouts he had been given. Others admitted trying to comprehend the documents, but they too eventually gave up: I was looking at them [Pubco use cases], but I could not understand them. I would read through them, but it seemed like they were not making sense, and it just kept on    they [the requirements subteam] were sending out a lot of documents (a designer).

Dissatisfied with the graphic designers’ unwillingness to learn the strategists’ tools, Wendy decided to implement a new approach: Ask the strategists to build first-cut “wire frames” (essentially, black and white schematics of Web pages with no graphics) and then give these to the designers. Although the strategists supported this decision, they believed the task was not their responsibility. They expressed frustration at seeing the designers go home at 6:00 pm while the strategists worked past the midnight to finish the wire frames. Nonetheless, as soon as the

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

designers received the first wire frame, they were up and running—working long hours and producing results. The information architect used the navigational choices made by the strategists in the wire frames to construct a site map for the new website. In fact, the designers were able to deliver several alternative graphic treatments for the index page in a few days. The project finally moved forward, after a month-long delay and only three weeks before the delivery date. Pubco’s core team members liked one of the designs and selected it to set the style for the rest of the mockup. In the evolution of the mockup, however, a new issue emerged. Initially, the graphic designers told the strategists that, for ease of use and aesthetic appeal, the top-level navigation should only have three to five navigation elements, which is the rule the strategists followed in the first set of wire frames they produced. However, as the wire frames were discussed with Pubco participants, new understandings of Pubco’s priorities were realized, leading to some changes to the index page navigation. Instead of three navigational elements, the index page soon had six. Later, when a seventh navigational element was added, the designers became increasingly frustrated with the sacrifices that had to be made in the design to accommodate the changes and wrote an e-mail to the group saying, “Speak now or forever hold your peace.” For example, designers had to “harden the edges” (make the banner and graphics look more square) and decrease the font sizes. When the requirements subteam decided on an eighth top-level navigational element, the design lost almost all of its “soft edges.” With a week left until the delivery date, the designers did not argue. “There was no time and the strategists did not want to talk to us,” one designer explained. For subsequent pages, strategists again produced wire frames, while designers added color and graphics. One of the designers reflected on the process: It helped getting wire frame forms. It—that process— even though it was like an assembly line, it helped as far as efficiency went. But    I was not understanding the basis behind it     I just remember thinking that if I had had the information [about what goes on a Web page] to begin with and had organized it, I definitely would have arranged it differently (a designer).

Pubco participants found the wire frames very helpful: “I cannot even imagine how you talk about what

123

is going to be on screen without using wire frames. I have got to use them” (Pubco requirements subteam member). At the end of the phase, however, neither Eserve nor Pubco participants particularly liked the outcome. The design? I’m pretty happy with it. But I do feel that it got watered down. Too many people    they were trying to accommodate all these different suggestions and comments and we’ve lost a bit of the originality of the original design. I actually prefer the very original. I really liked the light colors. I thought it was very sophisticated and soft (Pubco’s Web group manager, design subteam member).

The strategists blamed the graphic designers for not learning enough, not working hard enough, and not providing creative input into the wire frames. Wendy criticized the designer’s performance with their mentors at Eserve. The usability test conducted at phase completion indicated what designers knew all along—too many main navigational elements decreased the usability and aesthetic appeal of the site. Pubco decided to spend money on a new miniproject in which a team of Eserve-experienced graphic designers spent three more weeks on mockup redesign. In the implemented website there were just four top-level navigational elements and plenty of soft edges. CRIA Interpretation of the Prototyping Phase. CRIA interpretation of the interaction among the requirements and design subteams is depicted in Figure 5. The DILO document was supposed to be the boundary object par excellence for facilitating effective collaboration among strategists and graphic designers during the prototyping phase. It related directly to graphic designers’ practice by using a simple visual language. It was also concrete, given that it addressed the problem at hand—the website design (Bechky 2003). Moreover, it was accessible and current. However, it was not used for reflection by the graphic designers, who actually ignored it and were painfully stagnated in their work until the strategists produced the wire frames. Pubco use cases were produced through the challenging mode of CRIA enacted among Eserve strategists and Pubco participants (Figure 4). Yet these objects never became useful in communicating with graphic designers, who could not reflect on them.

124

Levina: Collaborating on Multiparty Information Systems Development Projects

The wire frames were actually among the most effective boundary objects in the prototyping phase in so far as they had acquired local usefulness and common identities across multiple boundaries. Requirements subteam members produced these objects by reflecting on Pubco use cases, DILO, and the existing website. They also went through several iterations of wire frames with Pubco participants and among themselves by engaging in the challenging mode of CRIA (Figure 4). Yet despite the usefulness of wire frames for enabling the graphic designers to reflect and for moving the design forward, the actual use of wire frames was such that the graphic designers only added to the designs produced by strategists. In essence, the mockup website was designed by strategists and only embellished with fonts and colors by graphic designers. The stakeholders were subsequently disappointed with the aesthetics and usability of the mockup, noting that the first versions of the design produced by the graphic designers were better than the final ones. Development of the Collective Identity and Power Negotiation The CRIA analysis helps address the first research question (understanding the collaborative process) and the first condition for achieving effective collaboration (the production of innovative, synergistic solutions). By sharing, reflecting, and using explicit objects, the participants engaged in a discourse across boundaries. In the Eserve-Pubco case, ignoring the expertise of others did not lead to synergistic solutions; adding led to some synergies; and only challenging enabled the transformation of diverse forms of expertise required for producing potentially innovative outcomes. The CRIA lens by itself, however, does not address the second research question of how diversity of agents’ backgrounds shapes the emergent design and the second condition for effective collaboration (balancing diverse interests). On the project, agents’ diverse backgrounds (practices and interests) were reflected in the IS design only if these backgrounds were first expressed in objects and second preserved through subsequent iterations of CRIA, i.e., either added to or challenged, but not ignored. Furthermore, participants’ degree of influence on the design was

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

greater if the adding, rather than challenging, mode was enacted in subsequent iterations of CRIA (see Figure 3). Agents had stakes in influencing the design: Each company’s position in its industry and individual professionals’ positions in their organizations and professional fields were tied to how the project unfolded. My analysis indicated that the mode of CRIA that was actually enacted depended on how agents used various sources of power on the project. There are many conceptualizations of power, often with contradictory assumptions and implications (see Hardy and Clegg 1996, Jasperson et al. 2002 for reviews). I draw on Bourdieu’s (1998) practice theory to conceptualize a source of power (or specie of capital) as a particular kind of relational resource that agents can accumulate and use to influence their own actions and those of other agents’ in a particular context.13 Moreover, “relational” here alludes to the notion that a particular resource may become a source of power in a relationship only if one agent has more of it than the other. According to Bourdieu there are three major species of capital to consider: economic (e.g., money or time), cultural/knowledge based (e.g., professional competence, education, ownership of information), or social (drawn from membership ties with a social group).14 Collective identities for groups arise from diverse participants taking a common interest in investing their resources into developing a new source of power (specie of capital). An advantage of drawing on this conceptualization of power is that power is seen as both inherited from external (often institutionalized) contexts as well as emergent in a given social context. Another advantage of this theory is that power is seen as both productive (through the combination and transformation of diverse species of capital, new 13

Space limitations prohibit me from fully developing the notion of power in this paper (see Levina 2001 for a more detailed treatment). Moreover, alternative conceptualizations of power might yield interesting insights into the use of the CRIA lens in future studies of collaboration.

14

In my data, economic and cultural resources were prominent sources of power; the social resources were less prominent. I will not discuss them in this paper. Moreover, symbolic capital, which is a special specie of capital that is used to claim any resource as valuable, is also not discussed here. See Levina and Vaast (2005) for a discussion of symbolic capital.

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

species of capital arise) and coercive (having capital allows one agent to take advantage of another). Agents’ diversity on the Eserve-Pubco project shaped which resources they could draw upon to influence collective actions and identity. Enacting ignoring or challenging modes of CRIA required that agents have sufficient capital to either completely dismiss or partially undo the collaborator’s influence on the design. Pubco Influencing Eserve in Shaping the Design. Control over payment terms (whether to hire Eserve and for how many phases) was a key economic capital that Pubco had during the project. When Pubco signed a narrowed-down contract stating that Eserve should rely on Pubco’s strategic documents to save money, it ignored Eserve’s methodology and implied that Eserve should engage more in the adding mode of CRIA in using Pubco’s prior work than in challenging it using Eserve’s expertise. Pubco used its economic capital (an ability to not fund the prototyping phase) in the production of the must-have list, thus ignoring Eserve’s prioritization methodology. Pubco also had cultural capital—such as the expertise in the book publishing industry and prior experience with the Web—and used it to challenge the design ideas expressed in use cases and wire frames. Eserve Influencing Pubco in Shaping the Design. Eserve participants also had economic levers as they decided on the level of investment they would make in a particular client’s project (Terwiesch and Loch 2004). When Pubco exercised its economic power to scale down the project, Eserve staffed the project with junior strategy consultants and shortened its duration. The compressed schedule, in turn, led to Eserve’s participants ignoring much (but not all) of Pubco’s prior work in the first few weeks of the project. Also, as a consultant, Eserve was in a position to use its cultural capital, namely its Web strategy and design expertise. Consulting research indicates that consultants are often in a position to abuse the power derived from their specialized expertise (Clegg et al. 2004, Elkjaer et al. 1991, Schön 1983, Yakura 1993). Consistent with this work, Eserve’s methods relied on specialized terminology that clients and even new Eserve consultants had trouble understanding. Eserve used this cultural capital to challenge Pubco’s prior

125

work in producing the sales presentation. It also continued to use its expert power in the planning phase to ignore or sometimes challenge Pubco’s prior work. However, because Eserve ignored so much of Pubco’s prior work, Pubco began suspecting that Eserve was abusing its expert power and subsequently used its economic capital to counteract Eserve’s cultural capital. Eserve Strategists Influencing Graphic Designers in Shaping the Design. Eserve strategists occupied key positions at Eserve, and most engagement and project managers were drawn from this discipline. There were no graphic designers in top management positions. As a result, strategists had greater control over the economic resources pertinent to projects. One of the ways in which they used these resources was to create tight schedules for their own and graphic designers’ deliverables on projects with the hope of generating higher profits on Eserve’s fixed-fee contracts. However, the designers’ work was highly nonlinear; many of them could go for days without producing anything while waiting for creative inspiration. Their training also encouraged them to evolve several designs at a time, knowing that only one would be implemented. Eserve’s compressed scheduled was partly responsible for why the designers on the Eserve-Pubco project only added to the wire frames produced by the strategists. Eserve’s practices also privileged strategists’ expertise (cultural capital) in Web design over designers’ expertise. Graphic designers were latecomers to Eserve. Those few who joined Eserve early typically had strategy, technology, and design education or experiences and were able to successfully contribute to the projects. As Eserve grew, however, two things happened: (1) Eserve methodology became established in a way that reflected the competence of the majority disciplines (strategy and technology), and (2) more “pure” graphic designers (not “Renaissance people”) were hired. Such practices as documenting IS designs in bullet lists (like the Pubco use cases) and delivering incremental results to clients were staples of management consulting practice but did not correspond to a graphic designer’s way of working. Strategists used this capital on the Eserve-Pubco project to produce Pubco use cases, which, on the one hand, helped them produce wire frames (boundary objects

126

Levina: Collaborating on Multiparty Information Systems Development Projects

that were used to facilitate collaboration) but on the other hand hurt the designers’ ability to challenge the IS design and thus inhibited effective collaboration. Moreover, the agents’ actual involvement in designing Pubco’s website became both a source of identity (commonality) for project participants and an emergent source of power for the strategists involved in the project. Eserve’s three-phase “waves” model, which became established when Eserve started growing and institutionalizing its practices, excluded designers from many early decisions about the site and limited their interaction with clients. Eserve strategists’ control over this source of power was used to ignore designers (when they asked strategists to stop adding navigational elements) as designers did not “get” Pubco’s business. Moreover, the designers lacked understanding of the client’s business to such a degree that they were simply unable to reflect on the DILO despite its “perfect” fit with graphic designers’ professional practices. This example demonstrates that ownership of some resources is necessary to enable even adding. Moreover, the project-based source of power is different from other sources of power used on the project in that it was not driven by some external differences among participants, but was produced through the project practices in a structurational manner. Understanding and influencing what has gone into shaping the design so far—what constituted the collective discourse and shaped the collective identity—became an important new specie of capital. It was related to, but not fully determined by, the actual time of joining the project (Wendy, for example, was able to learn what was behind the design through her organizational position and frequent communication with the clients and strategists). The more a particular agent influenced the design and the more his or her influence was preserved in subsequent iterations of CRIA (e.g., through the adding mode), the more project-based cultural capital the agent accumulated. Designers Influencing Strategists in Shaping the Design. Graphic designers did not have much control over economic capital at Eserve, but they did have a unique cultural capital—an ability to produce usable and aesthetically pleasing websites. Designers used this resource as a source of power for challenging the number of top-level navigational elements

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

depicted in wire frames (albeit, only to be ignored by the strategists). They also used this capital in ignoring Pubco use cases. They felt that had they invested in learning how to reflect on the bullet lists, they would have somehow strayed away from their professional identities: I did not read the majority of what she [a strategist] wrote. I am the absolute intended audience for this. These documents did not seem helpful to me     I am not a water faucet that just turns on and out comes beautiful art work. Beautiful artwork comes from taking a little time to create that beautiful art work. It is not just, you know, point and click, and here is beautiful art work. You have to know some things. And so, this has been a disappointment to me     But it is like, if I read all of her text, I am not inspired to do anything. I find it tedious     I just file it     I do not read that (graphic designer).

What we know and do as professionals defines who we are (Van Maanen and Barley 1984). Those few designers who had been with Eserve for a while and learned the strategists’ way of practicing moved into project leadership positions and, by and large, stopped practicing hands-on graphic design. Some designers recognized their disadvantaged positions at Eserve and even advised job candidates to seek design jobs elsewhere.

Discussion

To address the first research question, I drew on field study data and Schön’s theorizing about professional practice to argue that multiparty collaboration on ISD projects can be understood as a CRIA cycle. The cycle is enacted through the sharing of explicit objects. My analysis identified three different modes of the enactment of this cycle: ignoring, adding, and challenging. Usefulness of a particular object in enabling effective collaborative practices was not a property of the object per se; rather it was associated with ways in which the object was used in CRIA. The IS design emerged through the enactment of CRIA, which either preserved or transformed contributions to the design made by diverse agents. To achieve effective collaboration, at least part of the CRIA cycle needed to involve challenging. In addressing the second research question, I argued that the diversity in participants’ backgrounds exhibited itself in collaborative practice by forming

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

different sources of power, which in turn shaped which mode of CRIA was enacted. In addition to the sources of power associated with the control over economic resources, the power sources associated with control over cultural resources (e.g., expertise in Web strategy, publishing industry, and Web design) also played a critical role in shaping CRIA. Moreover, recognizing that an ability to shape the discourse and practices on the project was in itself a source of power provided unique insights into agents’ action. Who claimed control over this resource not only shaped CRIA, but also emerged from the CRIA cycle. To arrive at these conclusions I made some critical theoretical and methodological choices. My data-collection methods gave me an in-depth understanding of one specific case over a broader understanding of multiple cases. This choice was made because the phenomenon of interest (collaboration) unfolds through day-to-day interactions among ISD participants, is not fully understood by participants, and is important to study longitudinally. The specific practices that are described in this article (e.g., the use of wire frames) cannot be generalized beyond the Eserve-Pubco project, even to other Eserve projects. However, as argued by Lee and Baskerville (2003), theoretical constructs (such as CRIA, its modes, and the factors that shaped it) could be usefully applied to other settings. While readers familiar with ISD and consulting practice could argue that the dynamics of the project might be explained by Eserve’s poor project management skills, the consultants’ lack of experience, the designers’ laziness, or political interests and the manipulative moves of certain groups within Pubco, these explanations offer limited insights into collaborative dynamics on an ISD project. The concepts developed here can help us understand, for example, what is meant by good project management or consulting skills. Traditionally, project management is concerned with avoiding redundancy among developers’ efforts so as to achieve a joint goal (Kraut and Streeter 1995). Yet redundancy is a critical part of graphic designers’ practices (in one of Eserve’s other projects, more than 50 designs were produced, of which only one was chosen). Another effective practice on ISD projects is setting aggressive deadlines (Austin 2001). Yet this practice (implemented on

127

the Eserve-Pubco project) limits participants’ ability to produce innovative outcomes. Because the challenging mode of CRIA requires undoing some of what has already been accomplished, the manager needs to allocate time for this nonlinear process. Also, good consulting skills on collaborative projects involve reflecting on and finding ways of convincingly challenging clients’ prior work (Handley et al. 2005).

Implications

A persistent conclusion in IS literature is that more in-depth studies of ISD processes are needed to achieve a richer understanding of this complex social phenomenon (Newman and Robey 1992, Robey 1994, Sambamurthy and Kirsch 2000). I took this call seriously and developed a new CRIA lens for understanding multiparty collaborative practices in ISD. The new framing offers insights into challenges encountered on Web-based projects as marketing and design professionals enter these settings as representatives of consumers’ views. This study revealed that graphic design professionals have a harder time collaborating effectively in ISD settings that are dominated by business analysts. The time pressure on these projects also exacerbates this challenge, as designers tend to work in nonlinear ways. The framework developed here provides insights into understanding these challenges for both researchers and practitioners. By developing a CRIA lens, I was able to give further support to the notion that explicit objects play a crucial role in collaborative efforts. My analysis indicated that focusing on the object itself provided insufficient insight into whether an object would be effectively used in practice. Some objects (e.g., the DILO) may be particularly well suited to becoming boundary objects for a given set of agents but are simply not used by other agents, rendering them obsolete. Moreover, the production and use of boundary objects, which are often expected to facilitate communication, may aid or inhibit effective collaboration, depending on their use, which I conceptualized through different modes of CRIA. Indeed, wire frames were very useful in communicating what strategists wanted from designers. Yet the designs they captured were left unchallenged and inhibited effective collaboration. While organizational theorists have

128

Levina: Collaborating on Multiparty Information Systems Development Projects

moved beyond studying physical properties of an object toward understanding its “social properties,” such as whether an object represents the differences among groups (Bechky 2003, Carlile 2002), my findings suggest that further insights could be gained if the research focus were to shift from an object to the practices surrounding the use of an object. Second, this research has important implications for studying collaboration in diverse teams. The literature on collaboration in diverse settings has traditionally been concerned with stable, well-understood kinds of diversity such as demographic, organizational, and functional diversity (Guinan et al. 1998, Hardy et al. 2005, Polzer et al. 2002). The application of the CRIA lens highlights the usefulness of the structurational, practice-based view of diversity, which focuses on the emergent nature of diversity and power. While the distribution of the initial sources of power shaped the CRIA cycle, the CRIA cycle also gave rise to a new source of power associated with the degree of involvement of an agent in shaping the IS design.15 While I limited my analysis to a few prominent sources of power and showed how they shaped CRIA, other types of diversity, such as gender and age, and the power derived from them, could be incorporated into the analysis in a similar way. Furthermore, alternative conceptualizations of the notion of power may yield additional insights. Third, the study illuminates the often-unspoken but frequent occurrence of ignoring in collaborative practice. The distinction between the adding and challenging modes of work is not new. In organizational literature, theorists have contrasted “efficiency” and “innovation” or “exploitation” and “exploration” (March 1991). However, organizational theory has not highlighted the third basic mode of joint practice— ignoring. We are very familiar with this mode of practice from our everyday lives. From pieces of junk mail to prior research articles on our topic of interest, we ignore much more than we reflect on. The data show instances of ignoring associated with being unaware of an object intended for reflection, being unable to physically access it, or lacking the competency necessary to interpret it. However, ignoring also 15

Bunderson (2003) has recently noted the importance of paying attention to project-based sources of power in diverse teams.

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

occurred when the object was “fully accessible” and interpretable, as was the case with the sales presentation created in the Eserve-Pubco business development phase. Moreover, ignoring should not simply be seen as “dysfunctional.” Graphic designers ignored Pubco use cases to reaffirm their professional identities. While this ignoring inhibited effective collaboration on the project, it is possible that it aided in maintaining Eserve’s expertise in graphic design over the longer term. This suggests that research on organizational learning and collaboration should not omit this mode of action from its analysis. In future work it would be interesting to analyze the effects of ignoring on furthering interests of different stakeholders, as well as to compare the case of intentional and unintentional ignoring (what Bowker 1997 termed “strategic forgetting”) and the case of ignoring to the case of organizational silencing (Morrison and Milliken 2000). Finally, the use of IT for sharing objects has an impact on ignoring and other modes of CRIA that should be better understood. Beyond these implications for studying collaboration in general, there are also several specific implications of this work for understanding ISD practice. First, it suggests that to achieve effective collaboration, it might make sense to involve agents in lower status organizational positions (e.g., graphic designers in IT firms) in the project early and let agents in higher positions challenge their designs by drawing on their institutional sources of power. Second, this research indicates that the use of structured analysis and design techniques can be problematic when they give technical and strategy expertise privilege over graphic design expertise. Third, in managing externally sourced projects, firms should reflect more deeply on the nature of the relationship and the associated modes of control (Kern and Willcocks 2001, Choudhury and Sabherwal 2003). If the goals behind the project depend on effective collaboration, clients may want to stay away from behavioral modes of control. To conclude, the CRIA lens helps us further open the black box of ISD, not just to see which social forces and interests influence the design, but also to see how the design emerges through the sharing of the explicit objects produced by participants. Together the CRIA

Levina: Collaborating on Multiparty Information Systems Development Projects Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

lens and the analysis of diverse sources of power on projects help us better understand why participants on multiparty ISD projects adopt one or the other mode of practice and end up producing particular IS designs. Acknowledgments

The author would like to thank the members of her dissertation committee, Wanda Orlikowski, Jeanne Ross, and Paul Carlile, for their invaluable guidance in conducting this work. Also, without the generous participation of people from the two companies (Eserve and Pubco), this research would not be possible. She also greatly appreciates the useful comments from Cynthia Beath, the senior editor, and the associate editor, and three anonymous reviewers, which helped in shaping this manuscript. Maryam Alavi, Michael Davern, Patrick Parnaby, Joseph Reagle, Lee Sproull, and Jon Turner were very helpful in providing detailed comments on the manuscript. Special thanks go to Virtual City Informatics, Ltd., and HiArts Design Group for their help with the figures. The research reported here is funded in part by the Center for Information Systems Research at the Massachusetts Institute of Technology.

References Agar, M. 1980. The Professional Stranger: An Informal Introduction to Ethnography. Academic Press, New York. Andersen, N. E. 1990. Professional Systems Development: Experience, Ideas, and Action. Prentice Hall, New York. Austin, R. 2001. The effects of time pressure on quality in software development: An agency model. Inform. Systems Res. 12(2) 195– 207. Barki, H., J. Hartwick. 2001. Interpersonal conflict and its management in information system development. MIS Quart. 25(2) 195–228. Barley, S. R. 1990. Images of imaging: Notes on doing longitudinal field work. Organ. Sci. 1(1) 220–247. Bechky, B. A. 1999. Crossing occupational boundaries: Communication and learning on a production floor. Unpublished doctoral dissertation, Stanford University, Stanford, CA. Bechky, B. A. 2003. Sharing meaning across occupational communities: The transformation of understanding on a product floor. Organ. Sci. 14(3) 312–330. Bødker, S. 1998. Understanding representation in design. HumanComput. Interaction 13(2) 107–125. Bødker, S. 2000. Scenarios in user-centred design—Setting the stage for reflection and action. Interacting Comput. 13(1) 61–75. Bourdieu, P. 1977. Outline of a Theory of Practice. Cambridge University Press, Cambridge, NY. Bourdieu, P. 1998. Practical Reason: On the Theory of Action. Stanford University Press, Stanford, CA. Bowker, G. 1997. Lest we remember: Organizational forgetting and the production of knowledge. Accounting, Management Inform. Tech. 7(3) 113–138.

129

Bowker, G., S. Timmermans, S. L. Star. 1996. Infrastructure and organizational transformations: Classifying nurses’ work. W. J. Orlikowski, G. Walsham, M. R. Jones, J. I. DeGross, eds. Information Technology and Changes in Organizational Work. Chapman and Hall, London, UK, 344–369. Briers, M., W. F. Chua. 2001. The role of actor-networks and boundary objects in management accounting change: A field study of an implementation of activity-based costing. Accounting, Organ. Soc. 26(3) 237–269. Bunderson, J. S. 2003. Team member functional background and involvement in management teams: Direct effects and the moderating role of power centralization. Acad. Management J. 46(4) 458–474. Byrd, T. A., K. L. Cossick, R. W. Zmud. 1992. A synthesis of research on requirements analysis and knowledge acquisition techniques. MIS Quart. 16(1) 117–138. Carlile, P. R. 1997. Understanding knowledge transformation in product development: Making knowledge manifest through boundary objects. Unpublished doctoral dissertation, University of Michigan, Ann Arbor, MI. Carlile, P. R. 2002. A pragmatic view of knowledge and boundaries: Boundary objects in new product development. Organ. Sci. 13(4) 442–455. Carlile, P. R. 2004. Transferring, translating, and transforming: An integrative framework for managing knowledge across boundaries. Organ. Sci. 15(5) 555–568. Choudhury, V., R. Sabherwal. 2003. Portfolios of control in outsourced software development projects. Inform. Systems Res. 14(2) 291–314. Clegg, S. R., M. Kornberger, C. Rhodes. 2004. Noise, parasites and translation—Theory and practice in management consulting. Management Learning 35(1) 31–44. Elkjaer, B., P. Flensburg, J. Mouritsen, H. Willmott. 1991. The commodification of expertise: The case of systems development consulting. Accounting, Management Inform. Tech. 1(2) 139–156. Epstein, J. 2001. Book Business: Publishing Past, Present, and Future. W. W. Norton, New York. Faraj, S., L. Sproull. 2000. Coordinating expertise in software development teams. Management Sci. 46(12) 1554–1568. Gable, G. G., W. W. Chin. 2001. Client versus consultant influence on client involvement in computer system selection projects: A two-actor model of the theory of planned behavior. S. Sarkar, V. Storey, J. I. DeGross, eds. Proc. Twenty-Second Internat. Conf. Inform. Systems. New Orleans, LA, 249–260. Giddens, A. 1984. The Constitution of Society: Outline of the Theory of Structuration. University of California Press, Berkeley, CA. Glaser, B. G., A. L. Strauss. 1967. The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine Pub. Co., Chicago, IL. Guinan, P., S. Cooprider, S. Faraj. 1998. Enabling software development team performance during requirement definition: A behavioral versus technical approach. Inform. Systems Res. 9 101–125. Handley, K., A. Sturdy, T. Clark, R. Fincham. 2005. Client-consultant relations and the role of challenge. S. Newell, ed. Proc. Organ. Knowledge, Learning and Capabilities. Boston, MA. Hardy, C., S. R. Clegg. 1996. Some dare call it power. S. Clegg, C. Hardy, W. R. Nord, eds. Handbook of Organization Studies. Sage Publications, London, UK. Hardy, C., T. B. Lawrence, D. Grant. 2005. Discourse and collaboration: The role of conversations and collective identity. Acad. Management Rev. 30(1) 58–77.

130

Levina: Collaborating on Multiparty Information Systems Development Projects

Henderson, K. 1991. Flexible sketches and inflexible data bases: Visual communication conscription devices, and boundary objects in design engineering. Sci. Tech., Human Value 16(4) 448– 473. Hoegl, M., H. G. Gemuenden. 2001. Teamwork quality and the success of innovative projects: A theoretical concept and empirical evidence. Organ. Sci. 12(4) 435. Iivari, J., R. A. Hirschheim, H. Klein. 1998. A paradigmatic analysis of contrasting information systems development approaches and methodologies. Inform. Systems Res. 9(2) 164–193. Jacobson, I. 1993. Object-oriented Software Engineering: A Use Case Driven Approach. ACM Press, New York. Jasperson, J., T. A. Carte, C. S. Saunders, B. S. Butler, H. Cross, W. Zheng. 2002. Review: Power and information technology research: A metatriangulation review. MIS Quart. 26(4) 397–459. Kern, T., L. Willcocks. 2001. The Relationship Advantage: Information Technologies, Sourcing, and Management. Oxford University Press, UK. Kilker, J. 1999. Conflict on collaborative design teams: Understanding the role of social identities. IEEE Tech. Soc. Magazine 18(3) 12–21. Kim, J. Y., J. L. King. 2000. Boundary instances in heterogeneous engineering teams: Trouble management in the dram manufacturing process. Res. Managing Groups Teams 3 79–98. Kirsch, L., V. Sambamurthy, D.-G. Ko, R. Purvis. 2002. Controlling information systems development projects: The view from the client. Management Sci. 48(4) 484–498. Kraut, R. E., L. A. Streeter. 1995. Coordination in software development. Comm. ACM 38(3) 69–81. Kyng, M. 1995. Representations of work: Making representations work. Comm. ACM 38(9) 46–56. Lee, A. S., R. L. Baskerville. 2003. Generalizing generalizability in information systems research. Inform. Systems Res. 14(3) 221–243. Leonard, D., W. C. Swap. 1999. When Sparks Fly: Igniting Creativity in Groups. Harvard Business School Press, Boston, MA. Levina, N. 2001. Multi-party information systems development: The challenge of cross-boundary collaboration. Unpublished doctoral dissertation, Massachusetts Institute of Technology, Cambridge, MA. Levina, N., E. Vaast. 2005. The emergence of boundary spanning competence in practice: Implications for information systems’ implementation and use. MIS Quart. 29(2) 335–363. Maguire, S., C. Hardy, T. B. Lawrence. 2004. Institutional entrepreneurship in emerging fields: HIV/AIDA treatment advocacy in Canada. Acad. Management J. 47(5) 657–679. March, J. G. 1991. Exploration and exploitation in organizational learning. Organ. Sci. 2 71–87. Merriam-Webster. 1998. Merriam-Webster’s Collegiate Dictionary, 10th ed. Merriam-Webster, Springfield, MA. Miles, M. B., A. M. Huberman. 1984. Qualitative Data Analysis: A Sourcebook of New Methods. Sage Publications, Beverly Hills, CA. Morrison, E. W., F. J. Milliken. 2000. Organizational silence: A barrier to change and development in a pluralistic world. Acad. Management Rev. 25(4) 706–725. Newman, M., F. Noble. 1990. User involvement as an interaction process: A case study. Inform. Systems Res. 1(1) 89–113.

Information Systems Research 16(2), pp. 109–130, © 2005 INFORMS

Newman, M., D. Robey. 1992. A social-process model of useranalyst relationships. MIS Quart. 16(2) 249–266. Orlikowski, W. J. 2000. Using technology and constituting structures: A practice lens for studying technology in organizations. Organ. Sci. 11(4) 404–428. Orlikowski, W. J. 2002. Knowing in practice: Enacting a collective capability in distributed organizing. Organ. Sci. 13(3) 249–273. Pawlowski, S. D., D. Robey. 2005. Bridging user organizations: Knowledge brokering and the work of information technology professionals. MIS Quart. 29(4) 645–672. Polzer, J. T., L. P. Milton, W. B. Swann. 2002. Capitalizing on diversity: Interpersonal congruence in small work groups. Admin. Sci. Quart. 47(2) 296–324. Reagans, R., E. W. Zuckerman. 2001. Networks, diversity, and productivity: The social capital of corporate R&D teams. Organ. Sci. 12(4) 502–517. Robey, D. 1994. Modeling interpersonal processes during systemdevelopment—Further thoughts and suggestions. Inform. Systems Res. 5(4) 439–445. Robey, D., L. Smith, L. Vijaysarathy. 1993. Perceptions of conflict and success in information systems development projects. J. Management Inform. Systems 10(1) 123–139. Sabherwal, R. 1999. The role of trust in outsourced IS development projects. Comm. ACM 42(2) 80–86. Sambamurthy, V., L. J. Kirsch. 2000. An integrative framework of the information systems development process. Decision Sci. 31(2) 391–411. Sarker, S., S. Sahay. 2003. Understanding virtual team development: An interpretive study. J. Association Inform. Systems 4 1–38. Schmidt, K. 1997. Of maps and scripts—The status of formal constructs in cooperative work. S. C. Hayne, W. Prinz, S. Augustin, M. Pendergast, K. Schmidt, eds. Proc. Internat. ACM SIGGROUP Conf. Supporting Group Work: The Integration Challenge. ACM Press, New York, 138–147. Schön, D. A. 1983. The Reflective Practitioner: How Professionals Think in Action. Basic Books, New York. Schrage, M. 2000. Serious Play : How the World’s Best Companies Simulate to Innovate. Harvard Business School Press, Boston, MA. Schwartzman, H. B. 1993. Ethnography in Organizations. Sage Publications, Newbury Park, CA. Star, S. L. 1989. The structure of ill-structured solutions: Boundary objects and heterogeneous distributed problem solving. M. Huhn, L. Gasser, eds. Readings in Distributed Artificial Intelligence. Morgan Kaufman, Menlo Park, CA, 37–54. Star, S. L., J. R. Griesemer. 1989. Institutional ecology, “translations” and boundary objects: Amateurs and professionals in Berkeley’s Museum of Vertebrate Zoology 1907–39. Soc. Studies Sci. 19 387–420. Suchman, L. 1995. Representations of work: Making work visible. Comm. ACM 38(9) 56–64. Terwiesch, C., C. H. Loch. 2004. Collaborative prototyping and the pricing of custom-designed products. Management Sci. 50(2) 145–158. Van Maanen, J. 1979. The fact of fiction in organizational ethnography. Admin. Sci. Quart. 24 539–550. Van Maanen, J., S. Barley. 1984. Occupational communities: Culture and control in organizations. L. L. Cummings, B. M. Straw, eds. Research in Organizational Behavior. JAI Press, Greenwich, CT, 287–365. Yakura, E. 1993. Practices of power: Meaning and legitimation in information technology consulting. Unpublished doctoral dissertation, Massachusetts Institute of Technology, Boston, MA.